Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844838 | P a g ePerformance comparison of AODV and DSR under variousnetwork conditionsPrince Yadav, Pankaj AudichyaNavneet Kumar, Sandeep Kumar Jaiswal,Poonam Sisodia G.L.N.A Institute of Technology, MathuraAbstractMobile ad hoc networks (MANET) areself-created and self organized by a collection ofmobile nodes, interconnected by multi-hopwireless paths in a strictly peer to peer fashion.Ad-Hoc routing protocols such as Dynamicsource routing (DSR), Ad-hoc on demanddistance vector (AODV) and Destinationsequence Vector (DSDV) has been proposed tosolve the routing problem in Ad-hoc networks.Extended research has been done in comparingthe different proposed ad-hoc routing protocolunder various network conditions. Packetdelivery ratio (PDR), routing overhead, pathoptimality, end to end delay and throughput aresome parameters commonly used in thecomparisons. This survey paper provides theanalyzed performance of the routing protocolunder the various conditions.Keywords- MANET , AODV, DSR, Throughput.I. INTRODUCTIONMobile ad-hoc network is collectionwireless mobile node dynamically forming atemporary network without a infrastructure orcentralised administration. The nodes are free tomove randomly and organize themselves arbitrarily;thus, the network’s wireless topology may changesrapidly and unpredictably. Mobile nodes that arewithin each other’s radio range communicatedirectly via wireless links while far apart rely onother nodes to relay messages as routers. In ad-hocnetwork each node acts both as a host and a routerwhich forwards a data intended for some node.Hence it is appropriate to call such network as“multi- hop wireless ad-hoc networks” .Routing Approaches in Mobile Ad Hoc Network In ad hoc networks, routes are mainly multihop because of the limited radio propagation rangesand topology change frequently and unpredictablysince each network host moves randomly.Therefore , routing is an integral part of ad hoccommunications. Routing is to find and maintain routebetween nodes in a dynamic topology withpossibly unidirectional links, using minimumresources .Classification of Routing Protocol in MANET1. Table-driven or Proactive Protocols: InTable routing protocols each node maintains one ormore tables containing routing information of everyother node in the network. Some of table drivenprotocols are: Distance sequence distance vector(DSDV), Wireless Routing Protocols (WRP),Clustered Gateway Switch Routing (CGSR) etc .2. On- demand or Reactive Protocols: Inreactive protocols routes are created when it isrequired. On-demand protocols, unlike table drivenprotocols ones, establish a route to destination whenthere is a demand for it, usually initiated by a sourcenode through discovery process with in a network.Some of On-demand protocols are: Dynamic SourceRouting (DSR), Ad-hoc On demand Distance Vector(AODV), Temporally Ordered Routing Algorithm(TORA), Associativity Based Routing (ABR) etc.Figure . 1. Classification of Routing Protocol
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844839 | P a g eTABLE I : COMPARISON OF TABLE-DRIVENAND ON-DEMAND PROTOCOLSTable-Driven On-DemandAvailabilityof RoutingInformationImmediatelyfrom routingtableAfter the routediscoveryRouteUpdatePeriodicAdvertisementsWhen requestedRoutingOverheadProportional tothe size of thenetworkregardless to thenetwork trafficProportional tothenumber ofcommunicatingnodes andincrease withincreased nodemobilityII . Dynamic Source RoutingDynamic Source Routing (DSR) is therouting protocol for wireless networks. It’s forms aroute on-demand when a transmitting computerrequest one. However, it uses source routing insteadof relying on the routing table at each intermediatedevice. There are two major phases for the protocol :(a) the route discovery process which is based onflooding and is used to dynamically discover newroutes, maintain them in node caches, (b) the routemaintenance process, periodically detects andnotifies network topology changes.Route Reply would only be generated if themessage has reached the intended destination node(route record which is initially contained in RouteRequest would be inserted into the Route Reply) .To return the Route Reply, the destinationnode must have a route to the source node. If theroute is in the Destination Nodes route cache, theroute would be used. Otherwise, the node willreverse the route based on the route record in theRoute Request message header (this requires that alllinks are symmetric). In the event of fataltransmission, the Route Maintenance Phase isinitiated whereby the Route Error packets aregenerated at a node. The erroneous hop will beremoved from the nodes route cache; all routescontaining the hop are truncated at that point. Again,the Route Discovery Phase is initiated to determinethe most viable route .(a) Building of the route record(b) Propagation of the route reply.Figure 2: DSR Route Discovery Process.Advantages and limitations of DSRThis protocol uses a reactive approachwhich eliminates the need to periodically flood thenetwork with table update messages which arerequired in a table-driven approach. Theintermediate nodes also utilize the route cacheinformation efficiently to reduce the controloverhead .The disadvantage of this protocol is that theroute maintenance mechanism does not locallyrepair a broken link. The connection setup delay ishigher than in table-driven protocols. Even thoughthe protocol performs well in static and low-mobilityenvironments, the performance degrades rapidlywith increasing mobility. Also, considerable routingoverhead is involved due to the source-routingmechanism employed in DSR. This routingoverhead is directly proportional to the path length.III. Ad hoc On Demand Distance VectorRouting (AODV)Ad hoc On-Demand Distance Vector(AODV) routing is a routing protocol for mobilead -hoc networks and other wireless ad-hocnetworks. . It is an on-demand and distance-vectorrouting protocol, meaning that a route is establishedby AODV from a destination only on demand.AODV is capable of both unicast andmulticast routing . It keeps these routes as long asthey are desirable by the sources. Additionally,AODV creates trees which connect multicast groupmembers .The trees are composed of the groupmembers and the nodes needed to connect themembers. The sequence numbers are used by AODVto ensure the freshness of routes. It is loop-free, self-starting, and scales to large numbers of mobile nodes.AODV defines three types of control messages forroute maintenance:RREQ- A route request message istransmitted by a node requiring a route to a node Asan optimization AODV uses an expanding ringtechnique when flooding these messages. EveryRREQ carries a time to live (TTL) value that statesfor how many hops this message should be
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844840 | P a g eforwarded. This value is set to a predefined value atthe first transmission and increased atretransmissions. Retransmissions occur if no repliesare received. Data packets waiting to be transmitted(i.e. the packets that initiated the RREQ). Everynode maintains two separate counters: a nodesequence number and a broadcast_ id.Figure 3: AODV Route Discovery Process.RREP- A route reply message is unicasted back tothe originator of a RREQ if the receiver is either thenode using the requested address, or it has a validroute to the requested address. The reason one canunicast the message back, is that every routeforwarding a RREQ caches a route back to theoriginator.RERR- Nodes monitor the link status of next hopsin active routes. When a link breakage in an activeroute is detected, a RERR message is used to notifyother nodes of the loss ofthe link. In order to enable this reporting mechanism,each node keeps a ―precursor list, containing theIP address for each its neighbours that are likely touse it as a next hop towards each destination .In Fig.3 A possible path for a route repliesif A wishes to find a route to J The aboveFigure3illustrates an AODV route lookup session.Node A wants to initiate traffic to node J for which ithas no route. A transmit of a RREQ has been done,which is flooded to all nodes in the network. Whenthis request is forwarded to J from H, J generates aRREP. This RREP is then unicasted back to Ausing the cached entries innodes H, G and D .Characteristics of AODV Unicast , Broadcast, and Multicastcommunication. On-demand route establishment with small delay. Multicast trees connecting group membersmaintained for lifetime of multicast group. Link breakages in active routes efficientlyrepaired. All routes are loop-free through use of sequencenumbers. Only keeps track of next hop for a route insteadof the entire route. Use of Sequence numbers to track accuracy ofinformation. Use of periodic HELLO messages to trackneighbours.Advantages and limitations of AODVThe main advantage of AODV protocol isthat routes are established on demand anddestination sequence numbers are used to find thelatest route to the destination. The connection setupdelay is less. The HELLO messages supportingthe routes maintenance are range-limited, so theydo not cause unnecessary overhead in the network.One of the disadvantages of this protocol isthat intermediate nodes can lead to inconsistentroutes if the source sequence number is very old andthe intermediate nodes have a higher but not thelatest destination sequence number, thereby havingstale entries.Also multiple Route Reply packets inresponse to a single Route Request packet can leadto heavy control overhead . Another disadvantage ofAODV is that the periodic beaconing leads tounnecessary bandwidth consumption .IV. Performance MetricsSome important performance metrics can beevaluated:-A. Packet delivery fraction: The ratio of the datapackets delivered to the destinations to thosegenerated by the CBR sources. Packets deliveredand packets lost are taking in to consideration.B. Throughput: There is two representations ofthroughput; one is the amount of data transferredover the period of time expressed in kilobits persecond (Kbps). The other is the packet deliverypercentage obtained from a ratio of the number ofdata packets sent and the number of data packetsreceived .C. Average end-to-end delay: This is the averagetime delay for data packets from the source node tothe destination node.D. Packet Delivery Ratio: This is the ratio ofnumber ofpackets received at the destination to the number ofpackets sent from the source. In other words,fraction of successfully received packets, whichsurvive while finding their destination, is called aspacket delivery ratio .V. EXPERIMENTAL ANALYSIS USINGSIMULATIONAs already outlined we have taken two On-demand (Reactive) routing protocols, namely Adhoc On-Demand Distance Vector Routing (AODV)and Dynamic Source Routing (DSR). The mobilitymodel used is Random waypoint mobility model
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844841 | P a g ebecause it models the random movement of themobile nodes. For all the simulations, the samemovement models were used, the number of trafficsources was fixed at 10, the maximum speed of thenodes was set to 20m/s and the simulation time wasvaried as 10s, 15s, and 20s.Scenario 1: In this scenario some parameters with aspecific value are considered. Those are as shown intable 2.Parameter ValueNumber of nodes 10Simulation Time 10 secPause Time 5msEnvironment Size 800x800Transmission Range 250 mTraffic Size CBR (Constant Bit Rate)Packet Size 512 bytesPacket Rate 5 packets/sMaximum Speed 20 m/sQueue Length 50Simulator ns-2.29Mobility Model Random WaypointAntenna Type OmnidirectionalTable 2: Scenario 1 for implementation of AODVand DSRFigure 4: A Screenshot of 10 nodes of AODV NAM– Network animatorFigure 5: A Screenshot of 10 nodes of DSR NAM –Network animatorFigure 6: X Graph of 10 seconds simulation time ofAODVThe Figure 6 shows the X graph of AODV. By theFigure we see that as the simulation start the packetreceived and packet loss is initially zero, becauseinitially there is no CBR connection and nodestaking their right place. As the CBR connectionsestablish between the nodes the number of packetreceived increases but no packet loss is there, itmeans allgenerated packets are being received by the nodes.But the packet loss increases substantially on thesimulation time increases. Finally the packetreceived is more than the packet loss.
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844842 | P a g eFigure 7: X Graph of 10 seconds simulation time ofDSRThe Figure 6.6 shows the X graph of DSR.By the Figure we see that as the simulation start thepacket received and packet loss is initially zero,because initially there is no CBR connection andnodes taking their right place. As the CBRconnections establish the number of packet lostincreases very much as compare to packet received.It shows that mostly generated packets are beingdropped by the nodes. But the packet loss decreasessubstantially on the simulation time increases, andnumber of packet received increases substantially onthe simulation time increases.Scenario 2: In this scenario some parameters with aspecific value are considered. Those are as shown intable 3Parameter ValueNumber of nodes 10Simulation Time 15 secPause Time 5msEnvironment Size 800x800Transmission Range 250 mTraffic Size CBR (Constant Bit Rate)Packet Size 512 bytesPacket Rate 5 packets/sMaximum Speed 20 m/sQueue Length 50Simulator ns-2.29Mobility Model Random WaypointAntenna Type OmnidirectionalTable 3: Scenario 2 for implementation of AODVand DSRFigure 8: X Graph of 15 seconds simulation time ofAODVThe Figure 8 shows that the number ofpacket received increases according to simulationtime; it means generated packets are being receivedat a good ratio by the node. But the simulation timeincreases the packet loss increases substantially.Figure 9: X Graph of 15 seconds simulation time ofDSRFigure 9 shows that initially there is veryhigh packet loss but the number of packet receivedincreases according to simulation time; it meansgenerated packets are being received at a good ratioby the nodes. Because the simulation time increasesthe packet loss decreases substantially. TheScenario 3: In this scenario some parameters with aspecific value are considered. Those are as shown intable 4Parameter ValueNumber of nodes 10Simulation Time 20 secPause Time 5msEnvironment Size 800x800Transmission Range 250 mTraffic Size CBR (Constant Bit Rate)Packet Size 512 bytesPacket Rate 5 packets/sMaximum Speed 20 m/sQueue Length 50Simulator ns-2.29Mobility Model Random WaypointAntenna Type Omnidirectional
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844843 | P a g eTable 4: Scenario 2 for implementation of AODVand DSRFigure 10: X Graph of 20 seconds simulation timeof AODVThe above figures shows the same behaviorof AODV & DSR in fact of packet receiving andpacket loss, initially in AODV no packet loss and inDSR very high packet loss. But as simulation timeincreases the packet loss goes down and packetreceiving increases.Figure 11: X Graph of 20 seconds simulation timeof DSRVI. Conclusion and Future ScopeWe have compared two On-demand routingprotocols, namely, Ad hoc On-Demand DistanceVector Routing (AODV) and Dynamic SourceRouting (DSR). The simulation of these protocolshas been carried out using Ns-2 simulator on a―Pentium-IV, 2.0Ghz /RAM-1 GB /HDD- 120GB‖computer and ―Linux operating system.Three different simulation scenarios aregenerated and the simulation time has varied from10sec15sec and 20 sec. Other network parametersare kept constant during the simulation.It is observed that the packet loss is veryless in case of AODV, initially but it increasessubstantially on thesimulation time increases. In caseof DSR simulation the packet loss is very highinitially but it decreases substantially on thesimulation time increases.So, we can conclude that if the MANEThas to be setup for a small amount of time thenAODV should be prefer due to low initial packetloss and DSR should not be prefer tosetup a MANET for a small amount of time becauseinitially there is packet loss is very high. If we haveto use the MANET for a longer duration then boththe protocols can be used, because after some timesboth the protocols have same ratio of packetdelivering. But AODV have very good packetreceiving ratio in comparison to DSRThe two protocols Ad hoc On-DemandDistance Vector Routing (AODV) and DynamicSource Routing (DSR) have been compared usingsimulation, it would be interesting tonote the behaviour of these protocols on a real lifetest bed.In this work other network parameters suchas number of mobile nodes, traffic type-CBR,simulation area etc. are kept constant. Whereas thesimulation time is varied in the threedifferent simulation scenarios. It would beinteresting to observe the behaviour of these twoprotocols by varying these network parameters.REFERENCES Mandeep Kaur Gulati, Krishan Kumar -“AReview of QoS Routing Protocols onMANETs” 2013 Interntinal Conference onComputer Communication AndInformatics(ICCI-2013),Jan.04-06-2013,Combitore,INDIA. Ashish Bagwari, Raman Jee, Pankaj Joshiand Sourabh Bisht - “ Performance ofAODV Routing Protocol with increasingthe MANET Nodes and it’s effects on QoSof Mobile Ad hoc Networks”. 2012International Conference onCommunication Systems and NetworkTechnologies Rajneesh Kumar Gujral, Jitender Grover,Anjali, Sanjeev Rana.-” Impact ofTransmission Range and Mobility onRouting Protocols over Ad hoc Networks”.2012 International Conference onComputing Sciences. N. Adam, M. Y. Ismail and J. Abdulla,-“Effect of node density on performance ofthree MANET routing protocols”. ICEDSA-2010, pp. 321-325. K. Amjad and A. J. Stocker, “Impact ofnode density and mobility on performance
Prince Yadav, Pankaj Audichya / International Journal of Engineering Research andApplications (IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.838-844844 | P a g eof AODV and DSR in MANET”. CSNDSP-2010, pp. 61-65. Broch J., D. Johnson and D. Maltz, “TheDynamic Source Routing Protocol forMobile Ad Hoc Networks”,http://www.ietf.org/internet-drafts/draft-ietf-manet-dsr-03.txt, IETF Internet draft,Oct., 2009. David Maltz, Josh Broch, Jorjeta Jetcheva,and David Johnson.The effects of on-demand behavior in routing protocols formulti hop wireless ad hoc networks .IEEEJournal on Selected Areas inCommunication, 2009. Charles Perkins and Elizabeth Royer. Adhoc on-demand distance vector routing. InProceedings of the 2nd IEEE Workshop onMobile Computing Systems andApplications, pages 90–100, Feb2008. Charles Perkins, Elizabeth Royer, andSamir Das. Ad hoc on demand distancevector (AODV) routing. http://www.ietf.org/internet-drafts/ draft-ietf-manet-aodv-03.txt, June 2007. IETFInternet Draft (work in progress). Magnus Frodigh, Per Johansson and PeterLarsson ―Wireless ad hoc networking-Theart of networking without a network‖Ericsson Review No. 4, 2005. V.D Park and M.S Corson, ―A highlyadaptive distributed routing algorithm formobile wireless Networks‖,Proc.INFOCOM97, Apr.2005. J. Broch, David A. Maltz, David B.Johnson, Y-Ch Hu and J Jetcheva, ―Aperformance Comparison of Multi-hopWireless Ad Hoc Network for mobile adhoc networks‖, in Proc. 4th annualIEEE/ACM international conference onMobile Computing and Networking, Dallas,Texas, U.S. pp. 85-97, Oct. 2004.